A. The Sample of Firms

We focused on one Canadian province (British Columbia) and two American states (Washington and Georgia) that each had a number of comparable-technology pulp mills. In New Zealand and Australia, the number of similar kinds of mills was very much smaller and we could not choose particular subnational jurisdictions. A total of 15 mills agreed to cooperate, but one of the latter closed its doors, leaving us with 14.Footnote ⁵ We promised all our interviewees confidentiality in the write-up of information we obtained from them.

B. Qualitative Environmental Performance Data

At each facility, we conducted lengthy on-site, semi-structured interviews with environmental managers and, in most cases, mill managers. Our discussions were designed to elicit information about each facility's environmental management and pollution control history, control systems, challenges, approach to problems, and relations with regulators and environmental activists. We probed for specific examples, collecting detailed stories of particular environmental, regulatory, and economic problems, current or past, that illustrated the firm's characteristic response to challenges. In most cases, we also interviewed officials in corporate headquarters, regulators, and environmental activists familiar with the mills in question.Footnote ⁶ We sought additional perspective on each facility through interviews with industry association officials, environmental consultants, financial analysts, corporate lawyers, other commercial third parties, and, on occasion, mill employees.

C. Quantifiable Environmental Performance Data

Because good environmental performance requires progress on many dimensions, measuring relative success, even within an industry with comparable processes, is far from simple. In this article, we focus on water pollution alone, partly because of the elusiveness of comparable measures for the many types and sources of air pollution in pulp mills, and partly because of the consensus that certain contaminants in pulp mill wastewater, if not controlled, have especially serious impacts on aquatic ecosystems and human health. We concentrated, therefore, on obtaining the following measures of water pollution data, which pulp mills in all our jurisdictions are obligated to monitor and report.

1. 1 Biological Oxygen Demand (BOD), the standard measure of organic pollutant content of water, is a universally important measure of effluent quality. We were able to obtain 1998 and/or 1999 BOD data for 12 mills in the United States, Canada, and New Zealand (but not in Australia).Footnote ⁷

2. 2 Total Suspended Solids (TSS), the standard measure of particulate content of water, is another universally important measure of effluent quality. We were able to obtain 1998 and/or 1999 TSS data for 12 mills in the United States, Canada, and New Zealand (but not in Australia).

3. 3 AOX measures the level of absorbable organic halides in mills' effluent waters standardized by production level. AOX is a proxy measure for dioxins and furans, a family of persistent chlorinated organic compounds that accumulate in the food chain and have been associated with the poisoning of aquatic life, ecosystem damage, and possible human health effects. We were able to obtain comparable 1998 and/or 1999 AOX data for nine mills in the United States, Canada, and New Zealand (but not Australia).

In addition, for seven mills, we were able to obtain data on the incidence of accidental spills of chemicals used in the pulping and bleaching process. Such spills can result in toxic pollution and overwhelm the mills' wastewater-treatment systems, and they also are an indicator of the relative quality of the mills' environmental management program.

D. Assessing Environmental Management Style

In discussions of emerging strategies of environmental regulation, considerable emphasis has been placed on corporate environmental management systems, conceived of as formalized procedures for making and implementing corporate environmental policies, auditing results, and responding to shortcomings. Yet it has not been clear that adopting formalized management plans, which can be instituted for primarily symbolic reasons, actually produces improved environmental performance. Cary Reference Coglianese, Coglianese and NashCoglianese (2001), for example, has argued that managerial attitudes and actions—what he labels “commitment”—are the key variables in shaping corporate environmental performance. Our field research, accordingly, sought to identify “environmental management style” at each of the mills and corporations in our sample. By management style, we refer to the combination of managerial attitudes and actions that mark the intensity and character of each management's “commitment” to meeting environmental responsibilities. We focus on both attitudes and implementing actions because we learned that that managerial attitudes toward environmental matters could best be inferred from managers' accounts of the decisions and actions they had taken in response to particular regulatory or community challenges.

Thus we based our assessment of each mill's “environmental management style” on (1) managers' “expressed attitudes” toward environmental problems, (2) managers' actions and implementation efforts to meet specific economic, regulatory, and community challenges, and (3) their explanations for those actions. We relied on our interview data to score each firmFootnote ⁸ on three related dimensions of commitment to environmental values: (1) the intensity of managerial scanning for environmentally relevant information, including the search for “win-win” expenditures identified as both environmentally good and economically desirable for the firm; (2) the management's degree of responsiveness to environmentally relevant information, including demands from regulators, customers, neighbors, and environmental activists; and (3) the assiduousness with which the facility had institutionalized implementing routines to ensure high levels of environmental consciousness and control capacity (including activities such as self-auditing, employee training, and close integration of environmental and production-oriented training and decision making). We sought to increase the validity of our assessment of each firm on these dimensions by coding the qualitative data separately so that all three members of our research team agreed on the same characterization of the firm's environmental management style.

We found it difficult, however, to agree on a single numerical score for each mill because of the large number of activities within each of the three categories and because we could not easily weight the categories in terms of importance. However, we found ourselves in substantial agreement when we grouped the companies more holistically in terms of ideal types, while still drawing on the initially identified variables. We constructed five ideal types: Environmental Laggards, Reluctant Compliers, Committed Compliers, Environmental Strategists, and True Believers. Each successive managerial “type” displays incrementally greater commitment to compliance (or “overcompliance”) with regulatory requirements, scanning for environmental information and opportunities, responsiveness to regulators and environmental activists, and development of reliable implementing routines for their environmental policies. The Appendix contains a more extensive description of each type.

A. General Pattern: Compliance, Declining Pollution, Narrowing Differences

Over the last few decades, the pulp and paper industry has gradually reduced many important categories of environmentally harmful emissions. For the industry as a whole, biological oxygen demand (BOD) in effluent per ton of pulp produced decreased 90% between 1959 and 1988, total suspended solids (TSS) decreased by 80% between 1979 and 1988 (Reference Armstrong, Bentley, Galeano, Olszewski, Smith, Smith, Richards and PearsonArmstrong et al. 1998:123), and dioxin decreased by 90% between 1992 and 1998.Footnote ⁹ Canadian government data show that between 1990 and 1999, British Columbia's 22 pulp mills reduced BOD emissions (per ton of pulp produced) by 91%, TSS emissions by 50%, and AOX emissions by 83%, and there has been a considerable narrowing of differences between environmental “leaders and laggards” in levels of pollution.

Mill-level data from our sample facilities mimic that industrywide trend, as indicated by Figure 1, which displays total BOD emissions (as opposed to BOD per unit of pulp produced) for the two mills for which we obtained the longest time series data. In the late 1990s, for which we have good comparative data, JF was among the worst environmental performers in our sample, while TS was among the best; hence the trends for the other sampled mills probably have been quite similar.

Figure 1. Annual average BOD (lb/day) at TS and JF.

In 1999, none of the mills we studied were regulatory laggards in the sense of being ignorant of or systematic evaders of their “regulatory licenses.” All the mills in our sample generally were in compliance with their regulatory permits, and indeed their emissions of BOD and TSS generally were substantially lower than the levels required by their permits.Footnote ¹⁰ The finding of general compliance is in keeping with findings of other recent studies of the pulp and paper industry.Footnote ¹¹

Nevertheless, at the end of the twentieth century there remained significant differences among pulp and paper mills in environmental performance. Table 1 displays performance data among a sample of mills, averaged over 1998–1999, for three measures—BOD emissions (kg/day), TSS emissions (kg/day), and AOX emissions (kg/ton).Footnote ¹²

Table 1. Environmental Performance by Pulp Mill, 1998–1999

¹ This facility uses two different pulp production technologies onsite, one of which is far more polluting than the technology used at all the other facilities in our sample. The numbers shown here (for BOD and TSS) are figures constructed to estimate what pollutant discharges would have been if all production had been by the cleaner process. These are not the actual figures discharged by the facility.

The three lowest TSS dischargers in our sample (AK, SH, and TS) averaged 2,660 kg/day, only 34% as much as the three largest dischargers of TSS (AT, IK, and JF), which averaged 7,727 kg/day. The three lowest BOD dischargers (SH, PW, and TS) averaged 1,088 kg/day, which was only 23% as much as the three outliers (AT, IG, and JF, 4,769 kg/day). And, even setting aside the JF facility, which was a clear outlier in control of AOX effluent, the three lowest dischargers (RF, SH, and LV) pumped out only 36% as much AOX per ton of pulp produced as the three worst performers on that measure (TS, PW, and AK). The differences between leaders and laggards, it should be noted, is not because the “laggards” are larger mills or have higher levels of production than the leaders.Footnote ¹³

Although it is clear that reductions in these major pollutants have had positive and meaningful effects on the environment, the remaining effluent, even from the best performing pulp mills, continues to have negative environmental impacts. Thus we are left with an important puzzle. Why have some pulp mills done a better job in reducing pollution than others? Let us begin by examining the influence of different regulatory regimes.

B. Regulatory Regimes and Facility-Level Environmental Performance

Underlying the tradition of “command and control” regulatory laws, as well as much popular and academic writing about regulation, is a “deterrence model” of firm behavior, which holds that business firms abate environmental impacts only as required to by law and when they believe that noncompliance might be detected and penalized. In this model, variation in environmental performance depends on the interaction of: (1) the stringency of official environmental regulations; (2) the likelihood that violations will be detected (by officials or complainants); and (3) the severity of sanctions for noncompliance, as applied. From that standpoint, one would expect the best-performing pulp mills to be in the regulatory jurisdictions that have the most prescriptive and stringent legal rules, the most fearsome sanctions, and the most aggressive enforcement style. Other theories of regulation, however, argue that a uniformly aggressive style of regulation is likely to engender legalistic and political resistance, and that a more cooperative and flexible style of regulatory enforcement will generate higher levels of compliance, at least if regulators have the credible capacity to invoke strong legal sanctions against firms that fail to cooperate (Reference Bardach and KaganBardach & Kagan 1982; Reference ScholzScholz 1984; Reference Ayres and BraithwaiteAyres & Braithwaite 1992; Reference Gunningham and GrabowskyGunningham & Grabowsky 1998).Footnote ¹⁴ Whichever theory is correct, however, one would expect a significant correlation between regulatory style and environmental performance.

Prior research strongly indicates that the United States tends to employ a more prescriptive, deterrence-oriented style of regulation, whereas Canada, New Zealand, and Australia (like most other economically advanced democracies) generally employ a more cooperative, negotiated mode of enforcement.Footnote ¹⁵ Comparing enforcement of environmental regulations for pulp and paper mills in the United States and Canada, Kathryn Reference HarrisonHarrison (1995) concluded that the American regulations were enforced more legalistically, with more frequent penalties,Footnote ¹⁶ and that compliance with BOD effluent standards was higher in the United States. More impressionistic studies in Australia suggest that regulators there rely heavily on cooperation and persuasion, with an almost complete absence of the kind of coercive enforcement that often occurs in the United States (Reference GunninghamGunningham 1987; Reference Grabosky and BraithwaiteGrabosky & Braithwaite 1986; see also Reference SonnenfeldSonnenfeld 1996b). Moreover, compared to pulp mills in Canada, Australia, and New Zealand, mills in the United States are much more vulnerable to costly defense and potentially punitive private class action lawsuits for nonpermitted discharges that cause harm to fish, natural resources, or human health.

Although the level of regulatory prescriptiveness and legalism in American regulatory enforcement varies,Footnote ¹⁷ many environmental managers in the companies we visited reaffirmed the conventional differences in regulatory style. For example, corporate officials at a corporation with mills in both Canada and the United States contrasted the “enforcement frenzy” in the United States with the partnering approach regulatory officials took in Canada, where “We can stay focused on the end result and the regulators work with us on a compliance schedule. Often the important thing is keeping an eye on the goal. In the US they want the standard and they want compliance today.” Within the United States, some mill officials described U.S. EPA as more legalistic than state environmental enforcement officialsFootnote ¹⁸ and some American states, such as Georgia (where three mills in our sample are located), have long had the reputation of adopting a less legalistic approach to enforcement than some other states, such as Washington.

Nevertheless, our comparative data, dating from 1998–1999, found no consistent difference among regulatory jurisdictions in the environmental performance of “their” pulp mills, as shown by Figure 2. On none of the three measures (BOD, TSS, and AOX) do the facilities cluster tightly by regulatory jurisdiction, except for AOX emissions for mills in British Columbia, where regulations call for radical reductions in AOX by the end of 2002. Even so, two of the four BC pulp mills in our sample were slightly above the overall median in AOX emissions. Similarly, while two mills in BC had the lowest BOD and TSS emissions, two other BC mills were closer to the median, and one of those was slightly above it. Notwithstanding the supposedly greater deterrent threat of the American approach to regulation, the mills in the United States in 1998–1999 were as likely to be in the bottom half as in the top half of the environmental performance league. Moreover, the American mills in Washington (considered by some a politically “greener” state) did not do significantly better on average than those in Georgia. Indeed, variations among mills within each state were as large as differences across jurisdictions. More formally, when we compare the average emissions for each jurisdiction, there is no statistically significant difference across jurisdictions for BOD or AOX, and only British Columbia and New Zealand differed significantly on TSS.Footnote ¹⁹

Figure 2. Average emissions (1998–1999) for individual pulp mills, arrayed by regulatory jurisdiction.

Similarly, we failed to detect any significant statistical relationship between regulatory jurisdiction and the extent to which pulp mills had invested in state-of-the-art pollution control or reduction technology. We ranked facilities in terms on a 1–5 scale on a variety of control technologies.Footnote ²⁰ In addition to this “hardware-based” technology measure, we asked environmental managers to rank their own facility's control technologies from 1 to 5 (with 5 being the best available).Footnote ²¹ We found that regulatory regime does not predict technological sophistication; each jurisdiction has mills that are both above and below average on both hardware-based and subjective control technology measures.

The loose fit between regulatory jurisdiction and mill-level environmental performance in our data does not mean that differences among regulatory regimes do not matter at all. As noted above, BC's regulation requiring elimination of AOX emissions by the end of 2002—a requirement more stringent and more imminent than the U.S. cluster rule—helps explain the tighter clustering in AOX among Canadian mills. Similarly, U.S. regulations compelled pulp mills to install secondary wastewater-treatment facilities in the 1970s; but neither Canada nor British Columbia required secondary treatment for mills discharging into coastal waters until the 1990s. Thus, in 1998–1999, the time of the measures in Figure 2, Canadian coastal mills such as PW, SH, and RC had much newer (hence closer to “state of the art”) secondary treatment facilities than most American mills, and hence lower BOD and TSS emissions, on average.

Nevertheless, the overall weakness of the correlation between regulatory jurisdiction and firm-level environmental performance suggests (1) that at least by the mid-1990s, there was considerable convergence in the substantive pollution control standards in all our jurisdictions, and (2) within all the jurisdictions in this study, there is considerable variability in regulatory requirements, because (as indicated by Table 1) regulators tailor facility-level permits and informal orders to individual mills' inputs, technologies, surrounding environmental exigencies, and investment cycles (e.g., delaying stricter permit requirements for old facilities until a scheduled upgrade of its production processes). In no jurisdiction do regulations and regulators make all facilities march exactly together, as in close order drill; rather, like cowboys during a long, slow cattle drive, they prod a group of individuals in the same general direction. Differences in the details of regulatory permits may help explain some of the remaining differences in environmental performance between those at the front and at the back of the herd. But the correlation between mill-level regulatory permit limits and actual performance is only 0.46 for BOD, 0.30 for TSS, and those correlations did not reach statistical significance (based on nine facilities in Canada and the United States). Hence it is more likely that other factors, associated with the firms' economic situations, community pressures, and environmental management styles, are equally if not more important.

C. Economic Variables and Environmental Performance

Logically, firms that are more profitable and have greater financial depth should be able to sustain better environmental records than firms in the same industry that are pinched for profits and financially strained. Larger firms tend to be better environmental performers than smaller firms, partly because of their greater visibility and reputational concerns, partly because they have more resources to spare for specialized environmental engineering and management. For the most part, we could get financial data for only the corporations that own the individual mills, not for each pulp mill itself. Many of the parent corporations are vertically integrated, operating not only pulp mills (the subject of our research) but a variety of different kinds of paper mills and/or forestry divisions. We reasoned that although individual mills are expected to be financially independent to a considerable extent, they generally enjoy some degree of access to corporate financial resources for major capital investments. Moreover, since the mills' environmental failings might be attributed by the market or by regulators to the parent corporation, corporate-level resources presumably can be made available to deal with subpar environmental performance at the company's pulp mills if corporate officials desire or feel compelled to do so. Thus we tested the proposition that mills owned by larger corporations, and those with larger current profits and rising stock prices, would have better environmental improvements than those owned by corporations with lower sales, smaller (or negative) earnings, and declining share prices.

Our data provides some support for this proposition, provided that one assumes that it takes time for a corporation to translate economic good times into good environmental performance. First, we divided the mills in our sample into three categories—smaller, larger, and huge—based on the average annual sales of their corporate parents during the 1998–1999 period. We found no significant statistical relationship when we calculated average 1998–1999 BOD, TSS, and AOX emissions for each corporate size category. The same was true when we compared (1) corporate net income and (2) change in corporate stock price (up or down) with (3) their mills' environmental performance. Mills whose corporate parent was presumably experiencing milder economic constraints in 1998–1999 did not have lower BOD, TSS, or AOX levels than mills whose corporate parents were doing less well in that period.Footnote ²² Nor did we find a significant correlation between 1998–1999 corporate economic indicators and the relative 1998–1999 ranking of each pulp mill's environmental control technology (using either “objective” or “subjective” technology measures).

On the other hand, when we compared 1998–1999 environmental performance with corporate financial data in the 1990–1994 period, a different story emerged. Mills owned by corporations with larger profit margins (ratio of income to sales) and larger annual sales income in the early 1990s generally had lower BOD, TSS, and AOX emissions late in the decade, and also had better pollution control technology, although the relationships were not consistently significant for both measures of economic resources. See Table 2.

Table 2. Correlation Coefficients of Corporate Income and Income to Sales Ratio with Mill-Level Environmental Performance, Technology, and Management Style

¹ Only statistically significant at a p=0.10 level, all others statistically significant at a p=0.05 level (two-tailed). NS=not statistically significant.

Due to the importance of costly technology investments for reducing emissions,Footnote ²³ and the time lag for bringing new technology on line, it is plausible that corporate profitability at time 1 should have a stronger influence on environmental performance at time 2. Moreover, in our data, corporate profitability in the first half of the 1990s was correlated (0.62) with more ambitious environmental management style in 1998–1999, and as we will see later, environmental management style appears to have an independent influence on environmental performance.

Particular firm-level economic differences appear to be very important in explaining certain mills' relative environmental performance. Thus, due to variability in raw material and bleaching processes, some mills found investment in oxygen delignification systems to be profitable, enabling them to reduce emissions of AOX, on average, somewhat more than mills that concluded that oxygen delignification would not be cost effective.Footnote ²⁴

On the other hand, economic concerns constrain environmental managers in all firms when they push for nonincremental improvements. At least partly because of overcapacity in the world market for pulp and the weakness of customer demand for unbleached or totally chlorine-free paper, none of the mills in our sample had leapt far ahead of the others by abandoning pulp bleaching or by running a totally chlorine free (TCF) operation. One mill (SH) that tried TCF lost too much money and retreated.Footnote ²⁵ No companies in our sample had done the innovative engineering or made the very costly investments that would be necessary to operate a completely “closed loop” mill, with no discharges to surrounding waterways.Footnote ²⁶

D. Social Pressures

A growing body of literature focuses on the role of social pressures (or their absence) in shaping firm behavior vis-à-vis regulatory values. In some communities, environmental advocacy organizations, neighborhood groups, and local governments have intensified legal pressures by lodging complaints with regulatory agencies and courts (Reference Morag-LevineMorag-Levine 1994; Reference Sabatier and MazmanianSabatier & Mazmanian 1983; see also Reference SonnenfeldSonnenfeld 1996a, Reference Sonnenfeld1998a, Reference Sonnenfeld1998b). Environmental activists can intensify economic pressures by generating adverse publicity about polluting firms and organizing consumer boycotts. Mobilized communities also can have a direct political effect, inducing firms to improve environmental performance beyond the dictates of current law in hopes of forestalling tighter governmental regulation, or in hopes of preserving the local goodwill that facilitates smooth relations with local government and employees.

Virtually every pulp mill we visited reported significant environmental pressures from their host communities, which could hardly fail to notice the unpleasant odors and plumes of steam emitted by these facilities. With a constantly diversifying economy and a more mobile workforce, many mills that were once the only major employer in an isolated “company town” now are surrounded by more informed and more sophisticated communities in which they are not necessarily the dominant employer. Many mill managers spoke to us of having to meet the terms not only of their regulatory license but of their “social license.” As a mill manager told us, “We have to continuously convince the public we have a right to exist.” Social concerns about pulp effluent were intensified by a highly visible, worldwide campaign led by Greenpeace in the late 1980s and early 1990s that emphasized the threats to aquatic life and human health from the dioxin-laden effluent generated by the use of chlorine as a pulp bleaching agent (Reference SonnenfeldSonnenfeld 2002). In 1999, managers at one mill told us that the sanctions it feared most for breaching regulations were not legal but informal sanctions imposed by the public and the media, and hence it was motivated less by avoiding regulatory violations per se as “anything that could give you a bad name.”Footnote ²⁷

We could not generate quantifiable measures of the degree of social pressure on each mill, that is, of the relative “stringency” of the terms of each firm's social license. Activists demanded different things from different mills, and mills employed different ways of responding to, and hence reshaping, those demands. Thus we cannot compute a quantitative relationship between social license pressures and firm-level environmental performance.

A qualitative analysis, however, suggests that variations in social pressures have significant effects on firms' relative environmental performance. Consider again the contrast between TS and JF, as shown in Figures 1 and 2 and in Table 1. TS is among the leaders in control of BOD, and does reasonably well in controlling AOX, while JF is one of the comparative laggards. Yet both mills are in the same American jurisdiction, subject to the same federal and state laws, the same federal and state regulatory agencies, and hence presumably similar regulatory license requirements. While we do not have financial data for TS, part of a privately held corporation, it is a smaller corporation than JF, with less financial depth; in the latter half of the 1990s, JF had a better profit margin than most of the corporations in our sample for which public data was available. The two mills do differ significantly, however, in the immediacy and intensity of the community pressures they experience concerning environmental performance.

TS is located on the waterfront in a fairly large city, once heavily dependent on trade in forest products but now with a much more diversified economic base. TS's blue-collar and largely industrial neighborhood has changed; the TS mill is very visible to the white-collar workforce in nearby downtown office buildings and to commuters on a main roadway. The city is home to several environmental groups that pay close attention to TS's operations. When TS purchased its mill in the early 1980s from a company notorious for its lack of concern for the local environment while “shipping the profits back East,” the new owners found they faced an uphill battle in winning the trust of local activists and a local population who no longer wanted or needed a pulp mill on their doorstep. Yet the mill felt it needed to gain the community's trust. TS's environmental manager told us, “I see most of the [environmental] groups 2–3 times a month. Most important of all, what we say is what we do … I need their trust as much as they need ours. As soon as we violate it, all bets are off.”

TS's sensitivity to community pressure has at times prodded it to “beyond-compliance” measures. The company invited activists to all its permit negotiations with the regulators and it claims to have based its environmental priorities, at least in part, on those voiced by community activists. TS's sense of vulnerability to community action is conveyed by this account by its environmental manager.

JF, in contrast, is the primary employer in a small company town, miles from the nearest large city. It experiences some local pressures from owners of high-priced vacation homes whose riverside view encompasses the plumes from JF's pulp mill and some pressures from the government. After a chlorine release from the facility required a partial evacuation of JF's neighbors, the mill was obliged to install chlorine monitors both inside the mill and at the local fire department. Nevertheless, with its more isolated location, JF is not exposed to regular inquiries from local or regional environmental NGOs. JF managers do not voice the same sense of vulnerability to local pressures that TS's managers do. They have not come close to establishing the same level of communication or dialogue with NGOs or regulatory officials that TS's managers mention repeatedly. And unlike TS, they do not refer to voluntary “beyond-compliance” measures they have taken to win the trust of NGOs and regulators. Thus JF's less demanding “social license” appears to be a very important reason for the disparate environmental performance of the two mills.Footnote ²⁸

In several cases, customers' demands wedded environmental concerns with economic pressure, helping to explain certain performance differentials among the pulp mills in our sample. RF is the leader in our sample in eliminating AOX emissions, far exceeding the performance of the other American mills,which are subject to the same general standard for AOX in the EPA “cluster rule.” The company's decision to improve AOX performance was made in 1989 due to European customer concerns regarding dioxin in the diapers it produced. RF “Felt it's not good business to be in the middle of the [dioxin] controversy.”Footnote ²⁹

In 1998–1999, SH's environmental performance was excellent in all three effluent measures. A combination of environmental activism and consumer concerns a decade earlier helps explain why. In 1987, air pollution from SH was publicly criticized by news media, local environmental organizations, and local citizens (Reference Cashore and VertinskyCashore & Vertinsky 2000:14). In November 1988, dioxins were discovered in shellfish near SH's pulp mill in British Columbia. In highly dramatic public demonstrations, environmental activists targeted SH's parent company, a forestry company, and SH itself. The corporate CEO announced in 1988 that the firm would rebuild SH's facility, creating an environmentally friendly modern mill, and in April 1989 announced a new, wide-ranging environmental policy (Reference Cashore and VertinskyCashore & Vertinsky 2000:15). At the same time, environmental activists had been generating political pressures in British Columbia. In May 1989, the provincial environmental ministry announced tighter BOD and TSS limits, increased fines for violations, and in 1991 it demanded complete elimination of AOX discharge by December 31, 2002. Greenpeace kept up the pressure in the early 1990s; it organized a consumer boycott of chlorine-bleached paper in Germany, generating demands from some SH customers for chlorine-free paper. SH announced that it would “lead the way” in alternative bleaching processes and became the first company in North America to complete a full-scale mill trial of completely chlorine-free bleached softwood kraft market pulp (Reference Cashore and VertinskyCashore & Vertinsky 2000:16–17, 26; Reference RaizadaRaizada 1998; Reference WalshWalsh 1992). American pulp mills, on the other hand, generally are not export oriented and hence are more insulated from European demands for ECF and effluent-free paper (Reference Norbert-Bohm and RossiNorbert-Bohm & Rossi 1998:230). With the exception of RF, with its special customer concerns, the American mills in our sample had below-average records in AOX reduction in 1998–1999.Footnote ³⁰

As noted above, however, managers at different mills responded to social pressures in different ways. In some cases, our interviewees claimed that dramatic demonstrations against them by Greenpeace operated as a “wake-up call” that changed their corporation's attitudes toward its environmental responsibilities.Footnote ³¹ But it took a change in management before TS, despite its salience to the local community, adopted an open and receptive stance toward local activists. Mill AT, located in New Zealand, responded to Greenpeace's campaign against it by seeking court injunctions to restrain individual group members and engaging in its own media campaign to counter Greenpeace's arguments. As AT's environmental manager described it: “We decided to take the battle to Greenpeace, and our PR guy enjoyed the scrap. We decided we can win this war. We can visit the schools before they do, and build relationships with indigenous groups. It comes down to individuals and over time, to trust.” In summary, the company does not acknowledge any change in its behavior as resulting from the Greenpeace campaign and prefers simply to “tough it out.” (Note that in 1998–1999, AT had the weakest record in our sample for control of BOD and TSS emissions, although it was in the middle of the pack with respect to reduction of AOX emissions.)

E. Environmental Management Style and Environmental Performance

AT's response suggests that the influence of social pressures on environmental performance depends on an “intervening variable”—managerial attitudes. Our interviews, too, convinced us that firm-level “environmental management style” affects how firms respond to pressures from regulatory regimes and economic constraints. As noted earlier, we classified each mill's environmental management style on a scale extending from Environmental Laggard through Reluctant Complier, Committed Complier, Environmental Strategist, to True Believer.Footnote ³² We then correlated environmental management style with environmental performance.

The results are striking. As shown by Table 3, average BOD emissions for True Believers were substantially lower than those for Environmental Strategists, whose emissions were substantially lower than the average for Committed Compliers, whose emissions were substantially lower than the average for Reluctant Compliers. The same relationship emerges, albeit somewhat less dramatically, with respect to control of AOX and TSS (except that the Environmental Strategists did better than True Believers, on average, in controlling TSS). The correlation between environmental management style and environmental performance was 0.76 for BOD, 0.66 for TSS, and 0.57 for AOX. Thus in a cross-sectional analysis in 1998–1999, environmental management style was a much more powerful predictor of mill-level environmental performance than regulatory regime or corporate size and earnings.

Table 3. Management Style and Average Environmental Performance in 1998–1999

One reason that environmental management style is associated with better environmental performance is that True Believers and Environmental Strategists tend to invest in better pollution control technology. The correlation between management style and the “subjective technology” measure (based on managers' self-evaluation) is 0.67; with the “objective technology” measure (see note 19), it is 0.53. Higher-ranked “subjective technology,” in turn, is correlated with control of BOD (0.68), TSS (0.69), and AOX (0.8).Footnote ³³

But better technology is far from the only reason for better environmental performance by Environmental Strategists and True Believers. Much of their edge, our interviews suggest, stems more active “scanning” for “win-win” measures (which improve environmental performance and cut costs)Footnote ³⁴ and from a dedicated approach to day-to-day environmental management (what we have called “implementation”). Table 4 shows the relationship between mill environmental management style and the annual average incidence of chemical spills in 1998 and 1999 for seven U.S. mills. Two measures of spills are used: the first column shows the average number of all chemical spills reported to regulatory officials by each facility each year, the second the annual average of all “large spills.”Footnote ³⁵ Strikingly, the True Believers had perfect records, exceeding that of Environmental Strategists (who bested Committed Compliers on the first measure, but not the second).

Table 4. Management Style and Control of Chemical Spills, 1998–1999

Note: Numbers are average number of spills per facility per year. Big spills are those measuring more than 100 units, regardless of the unit of measure, because more hazardous materials tend to be measured in smaller units. Data available only for the seven U.S. facilities.

Environmental Strategists and, especially, True Believers also do a better job of building “reputational capital” with regulators and with environmental activists (in local communities and nationally), which appears to pay off in attaining more flexibility in regulatory permits.Footnote ³⁶

At the same time, environmental management style operates within important economic constraints. It is far from omnipotent in shaping environmental performance, and may well be shaped in part by the firm's economic situation. Two mills in our sample that were struggling economically through the 1990s nevertheless were True Believers in terms of environmental management style, and had good or excellent environmental performance scores. Nevertheless, we found that corporations that had larger sales and profit margins in the first half the 1990s were more likely to be Environmental Strategists in 1998–1999, while less profitable firms were more likely to be Committed Compliers. The quality of pollution control technology, too, is strongly correlated with economic performance, and it was the mills whose corporate parents did well financially in the 1990–1994 period that had the better environmental performance, on average, in 1998–1999. Finally, regulatory action and social pressures were the principal triggers for those expensive investments in pollution control technology, as visible ecological impacts and the environmental activism based on the dioxin scare in the late 1980s and early 1990s led to more stringent regulatory standards for AOX (and in Canada, for BOD and TSS).

We are left, therefore, with a complex, multivariate explanation for interfirm differences in environmental performance. Our sample is too small, unfortunately, to enable us to perform a multivariate regression analysis.

B. Understanding Convergence

Over time, there has been growing convergence across the countries and the firms we studied in the terms of each strand of the “license to operate.” Firms with operations in more than one jurisdiction did not regard their regulatory license as being materially different in different jurisdictions. Although they did refer to interjurisdictional differences of enforcement style and philosophy, they also observed that when the regulatory license ratcheted more tightly in one jurisdiction, other jurisdictions commonly followed that lead.Footnote ⁴⁰ Similarly, while many mills reported that they had experienced far less social pressure in an earlier era, all now experienced some such pressure, and communities tended to act as de facto regulators, thereby further diluting the importance of different enforcement styles. As one mill manager put it, “the implications of failing to meet the regulations are too great from a public or market point of view, so we are more demanding on ourselves than the regulators are.” Finally, an increasingly global and competitive world pulp market has diminished variability in the economic licenses of pulp mills. Institutional investors and financial analysts today are likely to judge all firms by common criteria.

Just as importantly, environmental demands among the different types of license have converged. Both regulatory and social licenses have demanded tighter controls on emissions. Moreover, economic and financial markets today are more prepared to look skeptically at firms that get adverse publicity for dramatic regulatory penalties or serious environmental accidents. Thus in our sample we did not find a single true Laggard, and only one full-blown Reluctant Complier. Simultaneously, however, a global-competition-driven economic license has in all jurisdictions constrained how far firms can go in a “green” direction. Thus when SH, as noted earlier, made a more far-reaching environmental investment in totally chlorine-free (TCF) technology than its industry peers, it ended up losing money, and the firms in our sample converged on the goal of substantially ECF operation. The net result of accommodating to the demands of the three different types of license is that a firm can neither afford to drop too low, nor aim too high: hence the considerable convergence in performance revealed by the statistics.

That convergence, however, has drifted toward better control of effluent in the pulp industry. The primary engine of that movement, we believe, has been periodic “tightenings” of governmental regulatory licenses. The law on the books (and in each mill's permit) is a benchmark for enforcers of both the social and economic license. Exposure of substantial legal noncompliance is taken by both community activists and professional investors as a justification for skepticism about the environmental good faith or the competence of mill managers. And that, of course, strengthens the capacity of regulatory license requirements to overcome economic license restraints. The largest reductions in pulp mill discharge to water of harmful pollutants have stemmed from investments in expensive control technologies, particularly secondary wastewater-treatment facilities and the substitution of chlorine dioxide for elemental chlorine as a bleaching agent (which often required construction of a chlorine dioxide plant). Economic license constraints often affected the timing of those installations, as firms often successfully argued that they should coincide with periodic rebuilding or updating of primary production equipment. But sooner or later, the regulatory license has trumped economic demands, partially through the implicit promise that all competitors would be obliged to make the same investment. And indeed, one of the most striking findings in our research has been the extent to which major investments in control technology have been made in response to pending or anticipated regulatory rules.Footnote ⁴¹

C. Understanding Variation

Notwithstanding substantial convergence in the terms of the various license requirements, significant variation between the licenses of different mills helped explain some of the differences in their environmental performance. With respect to the regulatory license, for example, we found that British Columbia's lag behind the United States in requiring secondary wastewater treatment in mills at the edge of coastal waters resulted in better BOD control by those BC mills in 1998–1999, on average, because their treatment facilities were newer and closer to “state of the art.” Similarly, BC's more imminent and more stringent regulatory deadline for elimination of AOX discharges helps explain why BC mills, on average, had lower AOX emissions in 1998–1999 than the American mills in our sample.

Other interfirm differences in environmental performance could be attributed directly to the terms of their particular economic license. A mill whose products (diapers) and customers (Western Europeans) were particularly environmentally sensitive had the lowest AOX discharges.Footnote ⁴² Some firms operating under serious economic license limits, such as those who were “cash-strapped,” told us this constrained their capacity to put in place appropriate environmental technology. Conversely, mills whose corporate parents had larger sales and higher profit margins in the first half of the 1990s, a period of intense social and regulatory pressures regarding chlorine, tended to have better technologies and better environmental performance at the end of that decade.

Different social license demands often appeared to be particularly powerful in influencing differences in environmental outcomes. For example, the gap between the emissions of JF and TS, described earlier, was very much what one would have anticipated, given JF's more remote, small-town location and TS's location near the heart of a changing, more economically diversified city with lively environmental activists.Footnote ⁴³ In a number of cases, our interview data suggested that a painful well-publicized encounter with a major environmental group produced a sea-change in the corporate approach to the environment.

D. Management Matters

Although firm behavior seemed clearly linked to the terms of its license to operate, we found a number of instances, described earlier, which could not be satisfactorily explained in these terms. Managers at different mills responded to social pressures in different ways. Corporate economic resources, as measured by sales and profitability, did not account for all the variance in mill-level environmental performance. And as noted, there usually was substantial variability in emissions by mills in the same regulatory jurisdiction. Accordingly, we hypothesized that the perceptions and attitudes of mill and corporate management, and their interpretation of their license terms, acted as an important filter through which information about the external licenses is sifted and guided their responsiveness to conflicting external pressures.

As described earlier, we found that there was indeed a statistically significant relationship between management style and environmental outcomes.Footnote ⁴⁴ Firms that we classified as Committed Compliers on the basis of our qualitative interview data had better control of their effluents, on average, than Reluctant Compliers, but the Committed Compliers did not do as well, on average, as the Environmental Strategists, who were in turn outperformed on most measures by the True Believers. In addition, if we compare the average and maximum BOD and TSS discharges of Environmental Strategists and True Believers (PG, VL, and TS) to that of Committed and Reluctant Compliers (IG, CB, JF) from 1990 through 1999, we find that all the former group show improvement in all parameters, and that this improvement is unambiguous and obvious in all but two (out of 12) cases. Among the latter group, only three cases (out of 12) of improvement occur, and in no case is the improvement obvious or unambiguous.

Yet firms still are constrained by the terms of their licenses. A firm that pushes the boundaries of its licenses too far will be punished: by regulators (if there is serious breach of the terms of a permit), by markets (if behavior goes beyond what is perceived by investors and analysts as economically rational), and by communities or NGOs (if its behavior goes far beyond what is perceived as socially acceptable). But since the limits on those licenses are unclear, and some of their terms are ambiguous, there is considerable scope for different environmental management to interpret them in different ways.

Although we are convinced both by our statistical and fieldwork data that “management matters,” our methodology did not enable us to explain precisely why firms approximated one ideal type or another. As a working model, we assume this will be the outcome of interaction between external factors (e.g., the license requirements), internal factors (e.g., corporate culture), and mediating factors (e.g., available corporate financial resources). Certainly, there was much to suggest that firms with different cultures behaved very differently. We were struck, for example, by the behavior of one “corporate raider” that operated two mills in our sample. In each case, its attitude to the local community was confrontationist in circumstances in which many other mills had gone to very substantial lengths to appease and establish trust with community groups. This attitude clearly came directly from the head office. The environmental manager at one of those mills (whose community liaison efforts had not received corporate support) summed it up as follows: “we've done a horrible job on PR in the last decade. There is a huge trust chasm with the local community … it's part of the corporate philosophy, to come in and take what you can get.” But to more fully tease out why different environmental management cultures arise, it would take a far more detailed and intensive study of a number of firms, including not just leaders (as a few studies have done)Footnote ⁴⁵ but also laggards (who are apt to refuse access to social scientists).